This is a perfect example to explain how to tackle such seemingly complex problems. In fact it is simple and straightforward:

First, we single out all lines with "TicketNumber===>" i them. These are the only lines we need to work on in this step:

Code:

sed -n '/TicketNumber===>/p' /path/to/inputfile | more

This does nothing more than to print the lines we want to work on: first check, if we found all the lines we want to find and didn'tfind all the lines we do not want to find. If the result is OK, we proceed.

Second, we catch the "word" immediately following the string "TicketNumber===>", because this is the ticket number itself. We display this ticket number instead of the original line to make sure we got that right:

Code:

sed -n '/TicketNumber===>/ {
              s/^.*TicketNumber===> \([^ ]*/) .*/\1/p
         }' /path/to/inputfile | more

Check again and compare with your input to make sure this is what you want - in case it isn't you will adapt the regexp until finally getting what you want.

Then proceed to the final step: we prepend every line with a ticket number in it with a field with the ticket number we have just isolated:

Code:

sed '/TicketNumber===>/ {
           s/^.*TicketNumber===> \([^ ]*/) .*$/\1:&/
      }' /path/to/inputfile > resultfile

Some obeservations: First, as you are only interested in the first and the last line of every transaction - that is, for every ticket number the opening and the closing line - you could throw away all lines in between, yes? If you search for both of these lines and prepend both with the ticket number a simple "sort" will provide the ordering so you can process the resulting file line by line:

Code:

sed -n '/TicketNumber===>/ {
              s/^.*TicketNumber===> \([^ ]*/) .*$/\1:&/p
         }
        /Task Set Response===> TT-/ {
              s/^.*Task Set Response===> TT-\([^ ]*/) .*$/\1:&/p
         }' /path/to/inputfile > resultfile

Second: you probably will not need some of the information in the source lines. By adapting the replacement part(s) of the regexps you can further trim down the resultfile to contain only the information you need.

You will probably have to fine-tune this to completely meet your needs, but this should give you a good start.

I hope this helps.

bakunin